Sunny Data designs and builds data platforms, model-backed workflows, optimisation systems, and scientific AI infrastructure for organisations where technical decisions have operational consequences.
Founder experience across
The work sits where data, AI, infrastructure, and software decisions are tightly coupled. Each challenge is framed around the operating reality first: users, ownership, failure modes, cost, governance, and maintainability.
Data platforms
When ingestion, modelling, and business rules drift apart, every downstream decision becomes fragile. The platform has to make ownership, lineage, quality, and cost visible by design.
Large-scale lakehouses · regulated environments · enterprise supply chains
AI systems
The model is rarely the hard part. Retrieval quality, permissions, latency, evaluation, retries, cost, and operational handoffs decide whether the system is usable.
Operational LLM workflows · retrieval and evaluation · enterprise systems
Machine learning
A useful model is only one part of the system. Feature logic, training, deployment, monitoring, drift response, and business ownership have to be designed together.
Risk modelling · scientific environments · production monitoring
Optimisation
Routing, scheduling, allocation, and planning work only when the constraints are modelled honestly. More code does not fix the wrong formulation.
Routing · resource allocation · constraint-heavy decisions
Research engineering
Research environments need speed without losing traceability. The system has to respect experimental workflows, reproducibility, and the realities of scientific teams.
Genomics platforms · oncology data · published biomedical work
Internal platforms
Dashboards and portals fail when they are treated as UI around weak data. The application has to be built from the operational data model outward.
Operational portals · public-sector workflows · backend and data ownership
Selected founder experience from enterprise platforms, regulated banking, global consumer systems, genomics infrastructure, and oncology research: different domains, same requirement: systems that survive real constraints.
Selected cases
01/05
Engagements are scoped around technical outcomes: a platform, a production workflow, an architecture decision, a migration, or a recovery plan. The work stays close to the people who own the system and the constraints that shape it.
Start from the current system: data, users, constraints, ownership, failure modes, and the decision that needs to be made.
The path is documented with alternatives, trade-offs, risks, and the reason it fits the client context before critical implementation starts.
Implementation focuses on the parts that determine whether the system can reach production: data contracts, infrastructure, workflows, interfaces, and operational behaviour.
Repositories, infrastructure, runbooks, and decision records are left in a state that the client team can reason about, operate, and evolve.
The difference is rarely the tool. It is how early the hard constraints are found, how clearly trade-offs are made, and whether the system can be owned after launch.
Unreliable data, latency, permissions, retries, drift, and handoffs are design inputs from the beginning.
Architecture decisions include constraints, alternatives, consequences, and the reason a path fits the client context.
Repositories, infrastructure, runbooks, and decision records are treated as part of the system, not project residue.
Readable systems, direct interfaces, and predictable operations usually matter more than impressive abstractions.
Sunny Data is an intentionally small technical engineering boutique based in Spain. Engagements are led directly by the founder, with a short path between technical diagnosis, architecture, implementation, and ownership.
The work sits at the intersection of enterprise data engineering, applied AI, optimisation, custom software, and scientific infrastructure: environments where technical ambiguity has real operational cost.
The operating model is direct: understand the constraints, design the system, build the critical path, document the reasoning, and leave the client with something they can maintain.
Background
Mathematician · Senior Technology Expert
20+ years building and operating production systems across telecommunications, entertainment, finance, and biomedical research.
IBM · Nokia · Microsoft · Mars · BCC
Technical range
The work crosses sectors, but the pattern is consistent: data, models, infrastructure, and decisions that have to survive operational reality.
Some systems need technical judgement before another implementation push. Sunny Data helps clarify what matters, what is risky, and what is worth building.